1. Feedforward neural networks initialization based on discriminant learning
- Author
-
Moncef Gabbouj, Kateryna Chumachenko, Alexandros Iosifidis, Tampere University, and Computing Sciences
- Subjects
Normalization (statistics) ,business.industry ,Computer science ,neural networks initialization ,Cognitive Neuroscience ,Initialization ,Pattern recognition ,113 Computer and information sciences ,Perceptron ,Convolutional neural network ,Unimodality ,Discriminant learning ,Machine Learning ,Neural networks initialization ,Artificial Intelligence ,discriminant learning ,Learning ,Feedforward neural network ,Neural Networks, Computer ,Artificial intelligence ,Projection (set theory) ,business ,Eigendecomposition of a matrix - Abstract
In this paper, a novel data-driven method for weight initialization of Multilayer Perceptrons and Convolutional Neural Networks based on discriminant learning is proposed. The approach relaxes some of the limitations of competing data-driven methods, including unimodality assumptions, limitations on the architectures related to limited maximal dimensionalities of the corresponding projection spaces, as well as limitations related to high computational requirements due to the need of eigendecomposition on high-dimensional data. We also consider assumptions of the method on the data and propose a way to account for them in a form of a new normalization layer. The experiments on three large-scale image datasets show improved accuracy of the trained models compared to competing random-based and data-driven weight initialization methods, as well as better convergence properties in certain cases. publishedVersion
- Published
- 2022